In many species, introduction of double-stranded RNA (dsRNA) induces potent and specific gene silencing. This phenomenon occurs in both plants and animals and has roles in viral defense and transposon silencing mechanisms. This phenomenon was originally described more than a decade ago by researchers working with the petunia flower. While trying to deepen the purple color of these flowers, Jorgensen et al. introduced a pigment-producing gene under the control of a powerful promoter. Instead of the expected deep purple color, many of the flowers appeared variegated or even white. Jorgensen named the observed phenomenon “cosuppression”, since the expression of both the introduced gene and the homologous endogenous gene was suppressed (Napoli et al., Plant Cell, 1990, 2, 279-289; Jorgensen et al., Plant Mol. Biol., 1996, 31, 957-973).
Cosuppression has since been found to occur in many species of plants, fungi, and has been particularly well characterized in Neurospora crassa, where it is known as “quelling” (Cogoni and Macino, Genes Dev. 2000, 10, 638-643; Guru, Nature, 2000, 404, 804-808).
The first evidence that dsRNA could lead to gene silencing in animals came from work in the nematode, Caenorhabditis elegans. In 1995, researchers Guo and Kemphues were attempting to use antisense RNA to shut down expression of the par-1 gene in order to assess its function. As expected, injection of the antisense RNA disrupted expression of par-1, but quizzically, injection of the sense-strand control also disrupted expression (Guo and Kempheus, Cell, 1995, 81, 611-620). This result was a puzzle until Fire et al. injected dsRNA (a mixture of both sense and antisense strands) into C. elegans. This injection resulted in much more efficient silencing than injection of either the sense or the antisense strands alone. Injection of just a few molecules of dsRNA per cell was sufficient to completely silence the homologous gene's expression. Furthermore, injection of dsRNA into the gut of the worm caused gene silencing not only throughout the worm, but also in first generation offspring (Fire et al., Nature, 1998, 391, 806-811).
The potency of this phenomenon led Timmons and Fire to explore the limits of the dsRNA effects by feeding nematodes bacteria that had been engineered to express dsRNA homologous to the C. elegans unc-22 gene. Surprisingly, these worms developed an unc-22 null-like phenotype (Timmons and Fire, Nature 1998, 395, 854; Timmons et al., Gene, 2001, 263, 103-112). Further work showed that soaking worms in dsRNA was also able to induce silencing (Tabara et al., Science, 1998, 282, 430-431). PCT publication WO 01/48183 discloses methods of inhibiting expression of a target gene in a nematode worm involving feeding to the worm a food organism which is capable of producing a double-stranded RNA structure having a nucleotide sequence substantially identical to a portion of the target gene following ingestion of the food organism by the nematode, or by introducing a DNA capable of producing the double-stranded RNA structure (Bogaert et al., 2001).
The posttranscriptional gene silencing defined in Caenorhabditis elegans resulting from exposure to double-stranded RNA (dsRNA) has since been designated as RNA interference (RNAi). This term has come to generalize all forms of gene silencing involving dsRNA leading to the sequence-specific reduction of endogenous targeted mRNA levels; unlike co-suppression, in which transgenic DNA leads to silencing of both the transgene and the endogenous gene. Introduction of exogenous double-stranded RNA (dsRNA) into Caenorhabditis elegans has been shown to specifically and potently disrupt the activity of genes containing homologous sequences. Montgomery et al. suggests that the primary interference effects of dsRNA are post-transcriptional; this conclusion being derived from examination of the primary DNA sequence after dsRNA-mediated interference a finding of no evidence of alterations followed by studies involving alteration of an upstream operon having no effect on the activity of its downstream gene. These results argue against an effect on initiation or elongation of transcription. Finally they observed by in situ hybridization, that dsRNA-mediated interference produced a substantial, although not complete, reduction in accumulation of nascent transcripts in the nucleus, while cytoplasmic accumulation of transcripts was virtually eliminated. These results indicate that the endogenous mRNA is the primary target for interference and suggest a mechanism that degrades the targeted mRNA before translation can occur. It was also found that this mechanism is not dependent on the SMG system, an mRNA surveillance system in C. elegans responsible for targeting and destroying aberrant messages. The authors further suggest a model of how dsRNA might function as a catalytic mechanism to target homologous mRNAs for degradation. (Montgomery et al., Proc. Natl. Acad. Sci. USA, 1998, 95, 15502-15507).
Recently, the development of a cell-free system from syncytial blastoderm Drosophila embryos that recapitulates many of the features of RNAi has been reported. The interference observed in this reaction is sequence specific, is promoted by dsRNA but not single-stranded RNA, functions by specific mRNA degradation, and requires a minimum length of dsRNA. Furthermore, preincubation of dsRNA potentiates its activity demonstrating that RNAi can be mediated by sequence-specific processes in soluble reactions (Tuschl et al., Genes Dev., 1999, 13, 3191-3197).
In subsequent experiments, Tuschl et al, using the Drosophila in vitro system, demonstrated that 21- and 22-nt RNA fragments are the sequence-specific mediators of RNAi. These fragments, which they termed short interfering RNAs (siRNAs) were shown to be generated by an RNase III-like processing reaction from long dsRNA. They also showed that chemically synthesized siRNA duplexes with overhanging 3′ ends mediate efficient target RNA cleavage in the Drosophila lysate, and that the cleavage site is located near the center of the region spanned by the guiding siRNA. In addition, they suggest that the direction of dsRNA processing determines whether sense or antisense target RNA can be cleaved by the siRNA-protein complex (Elbashir et al., Genes Dev., 2001, 15, 188-200). Further characterization of the suppression of expression of endogenous and heterologous genes caused by the 21-23 nucleotide siRNAs have been investigated in several mammalian cell lines, including human embryonic kidney (293) and HeLa cells (Elbashir et al., Nature, 2001, 411, 494-498).
Most recently, Tijsterman et al. have shown that, in fact, single-stranded RNA oligomers of antisense polarity can be potent inducers of gene silencing. As is the case for co-suppression, they showed that antisense RNAs act independently of the RNAi genes rde-1 and rde-4 but require the mutator/RNAi gene mut-7 and a putative DEAD box RNA helicase, mut-14. According to the authors, their data favor the hypothesis that gene silencing is accomplished by RNA primer extension using the mRNA as template, leading to dsRNA that is subsequently degraded suggesting that single-stranded RNA oligomers are ultimately responsible for the RNAi phenomenon (Tijsterman et al., Science, 2002, 295, 694-697).
Several recent publications have described the structural requirements for the dsRNA trigger required for RNAi activity. Recent reports have indicated that ideal dsRNA sequences are 21 nt in length containing 2 nt 3′-end overhangs (Elbashir et al, EMBO (2001), 20, 6877-6887, Sabine Brantl, Biochimica et Biophysica Acta, 2002, 1575, 15-25.) In this system, substitution of the 4 nucleosides from the 3′-end with 2′-deoxynucleosides has been demonstrated to not affect activity. On the other hand, substitution with 2′deoxynucleosides or 2′-OMe-nucleosides throughout the sequence (sense or antisense) was shown to be deleterious to RNAi activity.
Investigation of the structural requirements for RNA silencing in C. elegans has demonstrated modification of the internucleotide linkage (phosphorothioate) to not interfere with activity (Parrish et al., Molecular Cell, 2000, 6, 1077-1087.) It was also shown by Parrish et al., that chemical modification like 2′-amino or 5′-iodouridine are well tolerated in the sense strand but not the antisense strand of the dsRNA suggesting differing roles for the 2 strands in RNAi. Base modification such as guanine to inosine (where one hydrogen bond is lost) has been demonstrated to decrease RNAi activity independently of the position of the modification (sense or antisense). Some “position independent” loss of activity has been observed following the introduction of mismatches in the dsRNA trigger. Some types of modifications, for example introduction of sterically demanding bases such as 5-iodoU, have been shown to be deleterious to RNAi activity when positioned in the antisense strand, whereas modifications positioned in the sense strand were shown to be less detrimental to RNAi activity. As was the case for the 21 nt dsRNA sequences, RNA-DNA heteroduplexes did not serve as triggers for RNAi. However, dsRNA containing 2′-F-2′-deoxynucleosides appeared to be efficient in triggering RNAi response independent of the position (sense or antisense) of the 2′-F-2′-deoxynucleosides.
In one experiment the reduction of gene expression was studied using electroporated dsRNA and a 25 mer morpholino in post-implantation mouse embryos (Mellitzer et al., Mehanisms of Development, 2002, 118, 57-63). The morpholino oligomer did show activity but was not as effective as the dsRNA.
A number of PCT applications have recently been published that relate to the RNAi phenomenon. These include: PCT publication WO 00/44895; PCT publication WO 00/49035; PCT publication WO 00/63364; PCT publication WO 01/36641; PCT publication WO 01/36646; PCT publication WO 99/32619; PCT publication WO 00/44914; PCT publication WO 01/29058; and PCT publication WO 01/75164.
U.S. Pat. Nos. 5,898,031 and 6,107,094, each of which is commonly owned with this application and each of which is herein incorporated by reference, describe certain oligonucleotide having RNA like properties. When hybridized with RNA, these oligonucleotides serve as substrates for a dsRNase enzyme with resultant cleavage of the RNA by the enzyme.
In another recently published paper (Martinez et al., Cell, 2002, 110, 563-574) it was shown that double stranded as well as single stranded siRNA resides in the RNA-induced silencing complex (RISC) together with eIF2C1 and eIf2C2 (human GERp950 Argonaute proteins. The activity of 5′-phosphorylated single stranded siRNA was comparable to the double stranded siRNA in the system studied. In a related study, the inclusion of a 5′-phosphate moiety was shown to enhance activity of siRNA's in vivo in Drosophilia embryos (Boutla, et al., Curr. Biol., 2001, 11, 1776-1780). In another study, it was reported that the 5′-phosphate was required for siRNA function in human HeLa cells (Schwarz et al., Molecular Cell, 2002, 10, 537-548).
In one recently published paper the authors claim that inclusion of 2′-O-methyl groups into the sense, antisense or both the sense and antisense strands of a siRNA showed greatly reduced activity (Chiu, Ya-Lin and Rana, Tariq, M., RNA, 2003, 9, 1034-1048).
Recently, it has become clear that non-coding RNA genes produce functional RNA molecules with important roles in regulation of gene expression, as well as in developmental timing, viral surveillance, and immunity. Not only the classic transfer RNAs (tRNAs) and ribosomal RNAs (rRNAs), but also small nuclear RNAs (snRNAs), small nucleolar RNAs (snoRNAs), small interfering RNAs (siRNAs), tiny non-coding RNAs (tncRNAs) and microRNAs (miRNAs) are now known to act in diverse cellular processes such as chromosome maintenance, gene imprinting, pre-mRNA splicing, guiding RNA modifications, transcriptional regulation, and the control of mRNA translation (Eddy, Nature Reviews Genetics, 2001, 2, 919-929; Kawasaki and Taira, Nature, 2003, 423, 838-842). RNA-mediated processes are now also believed to direct heterochromatin formation, genome rearrangements, and DNA elimination (Cerutti, Trends in Genetics, 2003, 19, 39-46; Couzin, Science, 2002, 298, 2296-2297)
A large class of small non-coding RNAs known as microRNAs (miRNAs) plays a role in regulation of gene expression. In C. elegans, Drosophila, and humans, miRNAs are predicted to function as endogenous post-transcriptional gene regulators. The founding members of the miRNA family are transcribed by the C. elegans genes let-7 and lin-4, and were first dubbed “short temporal RNAs” (stRNAs). The let-7 and lin-4 miRNAs act as antisense translational repressors of messenger RNAs that encode proteins crucial to the heterochronic developmental timing pathway in nematode larva. (Ambros, Cell, 2001, 107, 823-826; Ambros et al., Current Biology, 2003, 13, 807-818).
Mature miRNAs originate from long endogenous primary transcripts (pri-miRNAs) that are often hundreds of nucleotides in length (Lee, et al., EMBO J., 2002, 21(17), 4663-4670). These pri-miRNAs are processed by a nuclear enzyme in the RNase III family known as Drosha, into approximately 70 nucleotide-long pre-miRNAs (also known as stem-loop, hairpin or foldback precursors) which are subsequently processed by the Dicer RNase into mature miRNAs (Lee, et al., Nature, 2003, 425, 415-419). The current model is that the pri-miRNA transcript is processed by Drosha in the nucleus, and the pre-miRNA hairpin precursor is exported from the nucleus and cleaved by Dicer in the cytosol to yield a double-stranded intermediate (similar in size to siRNAs), but only one strand of this short-lived intermediate accumulates as the mature miRNA (Ambros et al., RNA, 2003, 9, 277-279; Bartel and Bartel, Plant Physiology, 2003, 132, 709-717; Shi, Trends in Genetics, 2003, 19, 9-12).
miRNAs, like siRNAs, are processed by the Dicer enzyme, are approximately the same length, and possess the characteristic 5′-phosphate and 3′-hydroxyl termini. The miRNAs are also incorporated into a ribonucleoprotein complex, the miRNP, which is similar to the RNA-induced silencing complex, i.e., the RISC complex, through which siRNAs act (Bartel and Bartel, Plant Physiology, 2003, 132, 709-717). More than 200 different miRNAs have been identified in plants and animals (Ambros et al., Current Biology, 2003, 13, 807-818).
While siRNAs cause gene silencing by target RNA cleavage and degradation, miRNAs are believed to direct translational repression, primarily. This functional difference may be related to the fact that miRNAs tolerate multiple base pair mismatches whereas siRNAs are perfectly complementary to their target substrates (Ambros et al., Current Biology, 2003, 13, 807-818; Bartel and Bartel, Plant Physiology, 2003, 132, 709-717; Shi, Trends in Genetics, 2003, 19, 9-12).
Like the RNAse H pathway, the RNA interference pathway of antisense modulation of gene expression is an effective means for modulating the levels of specific gene products and may therefore prove to be uniquely useful in a number of therapeutic, diagnostic, and research applications involving gene silencing. The present invention therefore further provides compositions useful for modulating gene expression pathways, including those relying on an antisense mechanism of action such as RNA interference and dsRNA enzymes as well as non-antisense mechanisms. One having skill in the art, once armed with this disclosure will be able, without undue experimentation, to identify preferred compositions for these uses.